PostgreSQL - Insert data into multiple tables simultaneously - sql

Question

I am having a data insertion problem in tables linked by foreign key. I have read in some places that there is a "with" command that helps in these situations, but I do not quite understand how it is used.
I would like to put together four tables that will be used to make a record, however, that all the data were inserted at once, in a single query, and that they were associated with the last table, to facilitate future consultations. Here is the code for creating the tables:
CREATE TABLE participante
(
id serial NOT NULL,
nome character varying(56) NOT NULL,
CONSTRAINT participante_pkey PRIMARY KEY (id),
);
CREATE TABLE venda
(
id serial NOT NULL,
inicio date NOT NULL,
CONSTRAINT venda_pkey PRIMARY KEY (id)
);
CREATE TABLE item
(
id serial NOT NULL,
nome character varying(256) NOT NULL,
CONSTRAINT item_pkey PRIMARY KEY (id)
);
CREATE TABLE lances_vendas
(
id serial NOT NULL,
venda_id integer NOT NULL,
item_id integer NOT NULL,
participante_id integer NOT NULL,
valor numeric NOT NULL,
CONSTRAINT lance_vendas_pkey PRIMARY KEY (id),
CONSTRAINT lances_vendas_venda_id_fkey FOREIGN KEY (venda_id)
REFERENCES venda (id),
CONSTRAINT lances_vendas_item_id_fkey FOREIGN KEY (item_id)
REFERENCES item (id),
CONSTRAINT lances_vendas_participante_id_fkey FOREIGN KEY (participante_id)
REFERENCES participante (id)
);

Answer 1

The idea is to write WITH clauses that contain INSERT ... RETRUNING to return the generated keys. Then these “views for a single query” can be used to insert those keys into the referencing tables.
WITH par_key AS
(INSERT INTO participante (nome) VALUES ('Laurenz') RETURNING id),
ven_key AS
(INSERT INTO venda (inicio) VALUES (current_date) RETURNING id),
item_key AS
(INSERT INTO item (nome) VALUES ('thing') RETURNING id)
INSERT INTO lances_vendas (venda_id, item_id, participante_id, valor)
SELECT ven_key.id, item_key.id, par_key.id, numeric '3.1415'
FROM par_key, ven_key, item_key;

Answer 2

I know that you requested a single query, but you may still want to consider using a transaction:
BEGIN;
INSERT INTO participante (nome) VALUES ('Laurenz');
INSERT INTO venda (inicio) VALUES (current_date);
INSERT INTO item (nome) VALUES ('thing');
INSERT INTO lances_vendas (venda_id, item_id, participante_id, valer)
VALUES (currval('venda_id_seq'), currval('item_id_seq'), currval('participante_id_seq'), 3.1415);
COMMIT;
The transaction ensures that any new row in participante, venda and item leave the value of currval('X') unchanged.

Answer 3

You could create a function to do that job. Take a look at this example:
CREATE OR REPLACE FUNCTION import_test(p_title character varying, p_last_name character varying, p_first_name character varying, p_house_num integer, p_street character varying, p_zip_code character varying, p_city character varying, p_country character varying)
RETURNS integer
LANGUAGE plpgsql
AS
$body$
DECLARE
address_id uuid;
parent_id uuid;
ts timestamp;
BEGIN
address_id := uuid_generate_v4();
parent_id := uuid_generate_v4();
ts := current_timestamp;
insert into address (id, number, street, zip_code, city, country, date_created) values (address_id, p_house_num, p_street, p_zip_code, p_city, p_country, ts);
insert into person (id, title, last_name, first_name, home_address, date_created) values (parent_id, p_title, p_last_name, p_first_name, address_id, ts);
RETURN 0;
END;
$body$
VOLATILE
COST 100;
COMMIT;
Note how the generated UUID for the address (first insert) is used in the person-record (second insert)
Usage:
SELECT import_test('MR', 'MUSTERMANN', 'Peter', 'john2#doe.com', 54, 'rue du Soleil', '1234', 'Arlon', 'be');
SELECT import_test('MS', 'MUSTERMANN', 'Peter 2', 'peter2#yahoo.com', 55, 'rue de la Lune', '56789', 'Amnéville', 'fr');